Skip to main content

Full text of "SpinWaves in Ferromagnetic Dots 2D Honeycomb Lattice Stripes"

See other formats


VI 



Spin Waves in Ferromagnetic Dots 2D Honeycomb Lattice Stripes 

Maher Ahmed 

Department of Physics and Astronomy, 

University of Western Ontario, London ON N6A 3K7, Canada and 

! Physics Department, Faculty of Science, 

q Ain Shams University, Abbsai, Cairo, Egyp^\ 

(N 

Abstract 

In this work, the spin wave calculations were carried out using the Heisenberg Hamiltonian to study 
the allowed spin waves of zigzag and armchair edged stripes for ferromagnetic nanodots arrayed in a 2D 



honeycomb lattice [ Q. The Hamiltonian is used to construct the E matrix which encodes the exchange flow 
of magnons in the stripes. It is found that the allowed spin wave modes are the eigenvalues of the E matrix 



and therefore it is used to study the effects of the stripe width, edge exchange, the edge uniaxial anisotropy, 
and impurities on the allowed spin waves of stripes. The obtained results almost coincide with the results 
of graphene nanoribbons described by tight binding Hamiltonian for electronic excitations. Therefore, we 

s 

suggest the fabrication of the magnetic counterpart to graphene as a new technology in the field of spintronic 
devices and magnetic applications. 

O 

> 

in 
d 



> 

X 

S3 



* mahmed62@uwo.ca 



1 



I. INTRODUCTION 



The superior physical properties of graphene which lead to its promising applications in tech- 
nology are mainly attributed to both its crystal structure as 2D honeycomb lattice and its short 
range interactions. While graphene itself is not a strongly magnetic material, many experimental 
and theoretical works have been done for related magnetic properties and proposed designs of 
graphene-based spintronic device B2l|3l. Graphene is formed due to the nature bonding stability of 
carbon atoms, but there is no natural atomic elements able to form a stable planar ferromagnetic 
2D honeycomb lattice. In this case, however, ferromagnetic nanodots can be used as magnetic arti- 
ficial atoms [4] with the ability to design the requested magnetic properties in the same way as the 
quantum dots nanostructures are used as artificial atoms with ability to design tunable electronic 
properties not found in naturally existence atomic elements [HHEl- A question of interest here that 
arises to us from the physics and the technology point of view and from advances in the material 
science fabrication techniques is wherever are can fabricate a ferromagnetic dots 2D honeycomb 
lattice stripes as already done for similar magnetic structures ll9l-fTT1l. In this case, the ferromag- 
netic dots 2D honeycomb lattice stripes will share both crystal structure and short range interaction 
with graphene. The expected physical properties of that ferromagnetic dots 2D honeycomb stripes 
might lead to a new technology especially in the field of spintronic devices. A theoretical study is 
therefore needed to predict the similarity and the difference between magnetic and electronic short 
range interaction in the 2D honeycomb lattice. 

Many studies have been conducted on the thermodynamic properties of hexagonal and hon- 
eycomb atomic spin lattices lfT2l4T5ll . Many of these studies are for 3D systems reveal low- 
dimensional magnetic behavior with predominantly antiferromagnetic behavior like /?-Cu2V 2 07 
lfT6ll . or hexagonal spin lattices, with ferromagnetic and antiferromagnetic interaction |[T71l . Some 
studies have been conducted on the spin wave excitations in ferromagnetic nanostructures array 
lfT8l[T9l but there are no known studies for ferromagnetic nanodots 2D honeycomb lattices stripes 
which are not Bravais lattices. 

In this work, we will study the spin waves of zigzag and armchair stripes of ferromagnetic 
nanodots 2D honeycomb lattice with the total Hamiltonian ([T]), which was used in the study of 
spin waves of 2D square lattice in [|20ll . The new results will give us more understanding of lattice 
geometry effect on the physical properties of 2D materials. 

2 



II. THEORETICAL MODEL 



The systems under study here are 2D Heisenberg ferromagnetic stripes (or nanoribbons) formed 
from arrays of dots in the xy-plane with armchair and zigzag edges. We assume a honeycomb 
(graphene like) lattice with crystallographic description given in [21 J. The average spin alignment 
of the magnetic sites is in the z direction, which is also the direction of the applied magnetic field. 
The nanoribbon is of finite width in the y direction with N rows (labeled as n = 1, • • • , AT) and it is 
infinite in x direction (-00 <=> oo)(see Figure [T]). 




FIG. 1. Armchair (left) and zigzag (right) 2D Heisenberg ferromagnetic dots honeycomb stripes in xy- 
plane, where black (gray) dots are the sublattice A(B) with a line of impurities (white dots) in the middle of 
the sheet, and with average spin alignment in z direction. The stripes are finite in y direction with N rows 
(n = 1, • • • ,N) and they are infinite in the x direction. Figure taken from ll22l . 

The total Hamiltonian of the system is given by 

ij i i 

where the first term is the Heisenberg nearest-neighbor exchange term, the second term is the 
Zeeman energy term due to an applied field H Q , and the third term is the uniaxial anisotropy 
term. The summations over i and j run over all the sites where i and j always belong to different 
sublattice A(B). This is because because in the honeycomb lattice, the nearest neighbors of the A 
sites are always B sites and vice versa. The nearest neighbor exchange 7, ; has a constant "bulk" 
value J when either i and j are in the interior of the nanoribbons, and another constant value J e 
when i and j are both at the edge of the nanoribbon (i.e., in row n = 1 or n = N). Similarly, for 

3 



the site-dependent uniaxial anisotropy term £>,, we assume that it has a constant value D when the 
site i is inside the nanoribbon, and it is equal to D e for sites at the edge of the nanoribbon. 

To calculate the spin waves for this system at low temperatures T «. T c where the spins are 
nearly aligned such that S z - 5 for each spin, we use the Holstein-Primakoff (HP) transformation 
and follow the procedures of [QQ| to express the total Hamiltonian in terms of boson operators for 
the two sublattices A and B. We arrive to the expression 

#Total = E + H s (2) 
where the constant term E is the energy of the ground state for the ferromagnetic system given by 



E = S 2 



-?£j Ji >j-Tj Di -WbH J]s, (3) 



2 

V i,J i ) 



and the operator term H s has the following form 

Hs = " \ S E Ji 'j W + a < b l ~ b h ~ a ' ai ) (4) 
+ ^ [sMbHo + (25 - l)D k ] a] ai + Yj Is^bHo + (25 - 1)0*] b]bj, 

i j 

where a] (a,) and bj {bj) are the creation and the annihilation boson operators for sublattices A 
and B respectively. 

Since the nanoribbon extends to +oo in the x direction, we may introduce a ID Fourier trans- 
form to wavevector q x along the x direction for the boson operators a] (a,) and b ■ (bj) as follows: 

bj(x) = ^=J] b n (q x )e-^, b)(x) = -^=J] ^x)^, & 



di(x) = — != y a n (q x )e-^ r -, a\(x) = V al(q x )e^ r ' . 

' u n ' u n 

Here No is the (macroscopically large) number of spin sites in any row, q x is a wavevector in 
the first Brillouin zone of the reciprocal lattice and both r, and r ; is the position vectors of any 
magnetic sites i and j. The new operators obey the following commutation relations: 

[a n (q x ),al(q' x )]= 6 q ^ x , [b n (q x ), bl(q' x )] = 8^. (6) 

Also, we define the exchange sum: 

r(g x ) = \s J] J;,e '• '. (7) 



4 



The sum for the exchange terms 7,j is taken to be over all v nearest neighbors in the lattice which 
depends on the edge configuration as zigzag or armchair for the stripe. For the armchair configu- 
ration, the exchange sum gives the following amplitude factors y nn >(q x ) 



7nn'(qx) = \ SJ 

while for the zigzag case it gives 



exp( ic/ x -a)()„. M + exp ( i-q x a j <V. ; ,_.j 



(8) 



2 cos 



f V3 1 



3n'n±l + &n',n+l 



(9) 



The + sign depends on the sublattice since the sites line alternates from A and B. 

Substituting Equations ([5]) and ([7]) in Equation Q, and rewriting the summation over nearest 
neighbors sites, we get 



H s = 2] [a(ala nl + b\b n ^ + (y{q x )a n b\, + y(-q x )a\b n ^ . 



(10) 



q x ,nn' 



The first terms count the elementary excitations on each sublattice, while the second term describes 
the coupling between the sublattices, and a is defined by 



a = ( gf i B H + (2S -\)D n ) S n 



(11) 



In order to diagonalize H s and obtain the spin wave frequencies, we may consider the time 
evolution of the creation and the annihilation operators a] (a,) and b] (b f ), as calculated in the 

I J J 

Heisenberg picture. The equations of motion (using the units with fi = 1) for the annihilation 
operators ai{bj) are as follows Il23l - l27l : 



and 



(12) 



da n 

— = i[H,a n ] 
at 



i ^ -<*««' - y(-q x )b n < 



db n 
dt 



q x ,nn' 



= i[H,b n ] 



i 2] -ab n , - y(q x )a n , 



(13) 



q x ,nn' 



where the commutation relations between b] and bj in 

[b h fcj] = 6 tJ , [b], b] = -6 tJ , [b u b } ] = [b], bj] = 0. 



(14) 



was used and similar commutation relations between a] and aj, as well as the operator identity 
[AB, C] = A[B, C] + [A, C]B. 

The dispersion relations of the spin waves (i.e., energy or frequency versus wavevector) can 
now be obtained by solving the above operator equations of motion. The spin wave energy can be 
expressed in terms of the spin wave frequency using the relation E = fico, and assuming that spin 
wave modes behave like exp[-ico(q x )t]. We get the following sets of coupled equations: 



co(q x )a n = 2j ««n' + y(-q x )b n > 



q x ,nn' 



co(q x )b n = 2^ y(q x )a n ' + ab n * 

q x ,nn' 

The above equations can be written in matrix form as following 

0)(q x ) 



(15) 



(16) 



a n 




b„ 





b„ 



(17) 



det 



= 0. 



(18) 



aI N T(q x ) 
T*{q x ) aI N 

where the solution of this matrix equation is given by the condition 

-(a)(q x ) - a)I N T(q x ) 

T*(q x ) -(a)(q x ) - a)I N 

Here, T(q x ) is the exchange matrix, which depends on the orientation of the ribbon, and a)(q x ) are 
the energies of the spin wave modes. The matrix T(q x ) is given by 

' s B 

8 s y 

y e B 

B s 



(19) 



where the parameters e, y, and B depend on the stripe edge geometry and are given in Table [TJ 

Equation ( [T8] ) is very similar to the one obtained for graphene ribbons in reference ll22Tl . where 
the tight binding model was used. Where the next nearest-neighbor hopping term is neglected, 
the only essential difference between the Heisenberg model and the tight binding model results 
is the existence of the a term in the Heisenberg model. This extra term is in the diagonal of 
the Hamiltonian matrix which is shifting the total spin waves energy by amount related to the 
in-site Zeeman energy term and the uniaxial anisotropy energy term. This similarity between 



graphene and ferromagnetic stripes shows that Heisenberg and tight binding model are closely- 
related models with the nearest neighbor interactions represented by ty (Nearest neighbor hopping) 
and 



TABLE I. Nearest neighbor exchange matrix elements for 2D magnetic honeycomb lattice 
Parameter Zigzag Armchair 

E SJ- e -iq x a 

p 5/cos(V3^a/2) sj- e ic i* a/2 

SJ_ SJ_ iq x a/2 

y 2 2 e 



III. NUMERICAL CALCULATIONS 



The dispersion relations for the above 2D Heisenberg ferromagnetic dots honeycomb stripes are 



obtained numerically as the eigenvalues [28, 29 1 for the matrix Equation (17). The first step for 



solving this eigenvalue problem, for given value of the wavevector q x , is constructing the matrix 

cxIn T(q x ) 

E= W , (20) 

T*{q x ) aI N 



which is IN x IN since both I N and T(q x ) are N x N for the number N of rows in the stripe. 

First, the matrix aI N is independent of the value of the wavevector q x and it is simply con- 
structed using the material properties of the stripe for evaluating a values by Equation ( fTT| ), and 
since a is real the matrix q:In is also real. Second, the matrices T(q x ) and T*(q x ) depend on the 
value of the wavevector q x , the material properties of the stripe S J, and the stripe edge geometry 
as zigzag or armchair (see Table [[]). For the zigzag case the element of the exchange matrix T(q x ) 
are real (see Table [I]). Consequently T(q x ) = T*(q x ), and therefore the matrix E is real too. A 
standard procedure, following reference [29|, to obtain the eigenvalues for a real matrix is: 

First, balance the real matrix E by using similarity transformations in order to have comparable 
norms for corresponding rows and columns of the matrix, which then reducing the sensitivity of 
the eigenvalues to rounding errors. It is done here using the subroutine balanc Il29l . 

7 



Second, reduce the matrix E to a matrix that has zeros everywhere below the diagonal except 
for the first subdiagonal row, i.e., to upper Hessenberg form. It is done here using the subroutine 
elmhes (39). 

Third, find all eigenvalues of the matrix E in the upper Hessenberg form. It is done here using 
the subroutine hqr [|2"9~i 

Forth, sort the obtained eigenvalues of the matrix E (done here using the subroutine piksrt 
(29j) and plot the dispersion relations for the given stripe. 

For the armchair case, the element of the exchange matrix T(q x ) are complex (see Table [I]), 
and T(q x ) is Hermitian conjugate to T*(q x ), so consequently E is a Hermitian matrix. One way 
to obtain the eigenvalues of Hermitian complex matrix like E is to convert it to an equivalent real 
matrix 11291 , and then use the above standard procedures to obtain the eigenvalues for real matrix. 

The conversion to a real matrix is done as follows: First the Hermitian complex matrix E can 
be written as real and imaginary parts 



E = Re(q x ) + iJm(q x ) 



(21) 



where Re(g Y ) and Im(q x ) are IN x IN real matrixes, using the above representation of E in the 
Equation (fT7]), we get the following 2N x 2N complex eigenvalue problem 



a)(q x ) (u„ + i\ n ) = (Re(q x ) + ilm(q x )) ■ (u„ + i\ n ) 



(22) 



where u„ and y n represent the operators column vector. The above Equation ( |2"2"] ) is equivalent to 
solving the following AN x AN real eigenvalue problem 



to(q x ) 



u 




Re(^x) 


-Tm(q x ) 


u 


V 




1m(q x ) 


Re(q x ) 


V 



(23) 



A. Introducing effects of edges and impurities 

The study of edge effects on the 2D magnetic stripes properties are very important. They are 
introduced in this model numerically by using edges material properties in the elements (1,1) and 
(A 7 , A 7 ) in alu matrix, and the elements (1, 1), (1,2) and (2, 1) for lower edge and the elements 
(N, AO, (N, N - 1) and (N - l,N) for upper edge in the matrixes T(q x ) and T*(q x ). 

The pure 2D magnetic stripes offer very interesting dispersion relations, but to be suitable 
for technological devices applications one needs to engineer its properties. One way to do so is 



8 



the introduction of magnetic impurities lines substitutional^ into the stripe materials, which is 
very similar to suggested graphene materials engineering 11221 . The effects of one or two lines 
of impurities at any chosen rows numbers no and n' Q in the stripe are introduced in this model 
numerically by using magnetic impurities material properties in the elements (no, «o) an d (n' , n' ) in 
alu matrix, and in the elements of the matrixes T(q x ) and T*(q x ) that express the interaction of the 
impurity line with itself in the diagonal element (« , «o) f° r first line and (n' Q , n' Q ) for second line, the 
interaction of the impurity line with line before it in the stripe in the elements (n - 1, no), (n , n -l) 
for first line and (n' Q - 1 , n' Q ), (n' Q , n' - 1) for second line, and the interaction of the impurity line with 
line after it in the stripe in the elements (n , n +l),(n +l, n ) for first line and (n' Q , n' Q + 1), (n' Q +l, n' Q ) 
for second line. 

IV. RESULTS 

To compare our results for ferromagnetic dots stripes using the Heisenberg Hamiltonian with 
those of graphene nanoribbons using the tight-binding Hamiltonian, where the nearest neighbor 
(NN) interactions are represented by t t j and 7, 7 , we choose our stripes sizes, scaling our result to be 
dimensional less quantities, and choosing physical parameters matched that ones used in reference 
Il22ll for graphene. 

A. Zigzag stripes results 

Figure [2] shows the dispersion relation for zigzag 2D Heisenberg ferromagnetic dots honeycomb 
stripe with 20 lines without impurities, where the nearest neighbor exchange 7 i; has a constant 
value J through all the stripe including the stripe edges. The same goes for uniaxial anisotropy 
term £>,, which was chosen here to be zero such that the a is small and equal to 0.01. In this case, 
the Fermi level is 0.01, and the obtained dispersion relation is very near to the obtained dispersion 
relation zigzag graphene ribbon with same size [22J, as uniaxial anisotropy term D t increase to 1 
the a increase to 1.01. In this case, the Fermi level is 1.01, and all the dispersion curves shifted 
(see Figure [3]) as we discussed before about a effect. The figures [3] and [4] show that all modes 
have degeneracy of degree two which reflects the symmetry between the parallel rows of the two 
sublattices A and B in case of zigzag stripes. Also, the Figure [3] shows the famous localized edge 
states at Fermi level around q x = +0.57r/a for even zigzag graphene ribbons [|22] 1301 - 1321 . The 

9 



1 ' 1 ' 1 ' 1 ' 1 

-1.0 -0.5 0.0 0.5 1.0 

q a/ji 

^x 

FIG. 2. Spin waves dispersion for zigzag 2D Heisenberg ferromagnetic dots honeycomb stripes with N 
20, J = J e = 1, D = D e = and a = 0.01. 




-1.0 -0.5 0.0 0.5 1.0 

q a/jc 

FIG. 3. Spin waves dispersion for zigzag 2D Heisenberg ferromagnetic dots honeycomb stripes with ,/V 
20, J = J e = 1, D = D e = 1.0 and a = 1.01. 



10 



dispersion relation for even rows stripes is different from odd row (see Figure [4]) stripes especially 
for edges localized states at Fermi level. This is due to the fact that edges states depends on the 
probability of exchange between a site in the edge to an interior site. The two edges sites for even 
stripes have coordination number equal to 2 (i.e. each site at edges is connected to two interior 
sites of the stripe, where the stripe begins with sublattice A and ends with sublattice B (see Figure 
[T])). While the situation is different for the case of odd stripes, as one edge sites have coordination 
number equal to 2, while the other edge sites have coordination number equal to 1, as the odd 
stripe begins with sublattice A and ends with sublattice A (see Figure [T]). Which increases the 
localized edge states and extent it to fill the whole Brillouin zone for odd stripes. 

4- — ^ 



2- 




-1.0 -0.5 0.0 0.5 1.0 

q a/jt 

FIG. 4. Spin waves dispersion for zigzag 2D Heisenberg ferromagnetic dots honeycomb stripes with N = 
21, / = J e = 1, D = D e = 1 and or = 1.01. 

Figures [5] and [6] show the modified dispersion relations due to the effect of introducing substi- 
tutional a magnetic impurities line at row 1 1 of the zigzag stripes with 20 and 21 lines. The new 
dispersions show exactly the same behavior seen in the same case for zigzag graphene ||22"]j. but 
shifted in the case of magnetic stripes due to the effect of a. The introducing of the impurities line 
has the effect of splitting the stripe to two interacted stripes with different sizes. In case of 20 line 
stripe the new stripes are 10 lines and 9 lines, in case of 21 line stripe the new two stripes each 

11 



10 lines. The strength of the interaction between the two sub stripes depends on the value of the 
impurities exchange value Jj, the figures show case when Jj = 0. In this case, the expanded edge 
localized states in Fermi level are appear. Figure [6] shows the dispersions relation result for impuri- 
ties line at row 1 1 of 20 lines zigzag stripe as a superposition of dispersion relations for even stripe 
(here 10 lines) with odd stripe (here 9 lines), while Figure [5] shows the dispersions relation result 
for impurities line at row 11 of 21 lines zigzag stripe as a superposition of dispersion relations for 
two even stripe (here 10 lines) but shifted since the two stripes begin with different sublattice one 
A and the other B (23. 

Figures[7]and[8]show the modified dispersion relations due to the effect of introducing magnetic 
impurities lines at rows 1 1 and 14 of the zigzag stripes with 20 and 21 lines. Again, the introduction 
of impurity lines has the effect of splitting the stripe to three interacted stripes with different sizes; 
in case of 20 line stripe the new stripes are 10 lines, 2 lines and 6 lines, in case of 21 line stripe 
the new two stripes 10 lines, 2 lines and 7 lines. The existence of stripe with 2 lines between zero 
exchange lines creates accumulation sites for magnons which then create two fiat localized states: 
one in conduction band and the other in valance band. If the stripe size increases, the accumulation 
decreases and the localized states are removed. 

We see from the results above that zigzag type stripes are gapless even when impurities are 
introduced. To study the effects of edges and impurities on their dispersions relations, some pa- 
rameters are needed to reflect the important change in their dispersions relations due those effects. 
For the applications of zigzag graphene nanoribbons it has been found that the localized states at 
Fermi level are very important [|2T1 [33ll . Nakada et al.|[32||, used the density of states at Fermi level 
and at center band to study size effect on zigzag graphene ribbons. Here, we use the following two 
parameters to study the effects of edges and impurities on zigzag stripes: The first parameter is the 
relative density of states near the Fermi level (RDSFL), which computationally is calculated by 
finding the total number of points in the dispersion relations between a - 0.0002 and a + 0.0002, 
which is relative measure for the density of states at Fermi level. The second is the relative density 
of states of center band (RDSCB), which computationally is calculated by the total number of 
points in the dispersion relations between a - 1.0002 and a + 1.0002, which is relative measure 
for the density of states at center band. 

12 




FIG. 6. Spin waves dispersion for zigzag 2D Heisenberg ferromagnetic dots honeycomb stripes with an 
impurity line at line number 1 1, where N - 21, J - J e - I, Jj - 0.07 D = D e = Dj - 1.0 and a - 1.01. 



13 



1 ' 1 ' 1 1 1 ' 1 

-1.0 -0.5 0.0 0.5 1.0 

q a/n 

FIG. 7. Spin waves dispersion for zigzag 2D Heisenberg ferromagnetic dots honeycomb stripes with two 
impurities lines at line number 11 and line number 14, where N - 20, J - J e - I, Jj - 0.07, Jjj = 0.07, 
D = D e = Dj - \.0 and a = 1.01. 




FIG. 8. Spin waves dispersion for zigzag 2D Heisenberg ferromagnetic dots honeycomb stripes with two 
impurities lines at line number 11 and line number 14, where N = 21, 7 = J e - 1, 7/ = 0.07, Jjj = 0.07, 
D = D e = Dj = 1.0 and a = 1.01. 14 



1. The effect of zigzag stripe width on RDSFL and RDSCB 



To use the RDSFL and RDSCB as parameters to study the effects of edges and impurities on 
zigzag stripes dispersions relations, we need to make a calibration for those two parameters. We 
do that by studying the effect of zigzag stripes width on RDSFL and RDSCB, where edge and 
impurities effects are not taken in the account. 



200- 
180- 
160- 
140- 
120- 

§5 100 

Q 

CC 80- 
60- 
40 
20- 
0- 



20 



Odd Number N 



Even Number N 



40 



I 

60 



I 

80 



100 



N number of rows 



FIG. 9. The effect of zigzag stripe width on the relative density of states near Fermi level (RDSFL) 

Figure [9] shows the effect of zigzag stripe width on RDSFL, which is dependent on the stripe 
width parity as odd or even. In the case of odd stripe width, the RDSFL is constant and independent 
on the stripe width. This is understood from the large contributions of the localized edge states 
that extend over the entire Brillouin zone for edge sites with 1 coordination number. For even 
stripe width, the situation is different. The RDSFL is dependent on the stripe width, it is increase 
stepwise, where the step width increases with increasing stripe size. This could be explained by 
the fact that the probability of tunneling (or the diffusion length) for magnons between the two 
edges is high for small even width stripes and decreases with increasing even width stripes. This 
result for even width stipes is close to the results of Nakada et al.[32J for graphene zigzag ribbon, 
with the caveat that their zigzag line numbering for width always gave them eve zigzag stripes 
with our convention. 

15 




FIG. 10. The effect of zigzag stripe width on the relative density of states of center band (RDSCB) 



While Figure 10 shows the effect of zigzag stripe width on RDSCB. RDSCB is weakly depen- 
dent on stripe width parity as odd or even (i.e. the odd RDSCB is almost equal to the even RDSCB, 
save for a slight increase). This is understood since RDSCB includes RDSFL in its computation 
for both even and odd stripe width, while the center band almost the same for even and odd stripes. 
The RDSCB linearly increases with the stripe width for both even and odd width. 



2. The effects of edge uniaxial anisotropy on zigzag stripe RDSFL and RDSCB 



Figures [9] and 10 show RDSFL and RDSCB when the edge uniaxial anisotropy D e is equal to 



the stripe interior uniaxial anisotropy D. As D e becomes unequal to D, the RDSFL becomes zero 
for both even and odd stripes. This is understood since the edge localized states are a reflection 
of the decreasing probability of exchange interaction between the edge sites and interior sites, 
which in case of D e = D come from different coordination number for edge sites. At the same 
time, insite energy is equal between the edge and interior sites. As D e becomes unequal to D a 
difference in insite energy between the edge and interior sites is created. This difference in insite 
energy increases the probability of exchange interaction between the edge and interior sites which 
removes the localized states at the edge of zigzag stripes. 

16 




N number of rows 



FIG. 1 1. The effects of edge uniaxial anisotropy and zigzag stripe width on its RDSCB 



While RDSFL becomes zero as D e ± D, Figure 1 1 shows the color contour plot for the effects 
of edge uniaxial anisotropy and zigzag stripe width on its RDSCB. The figure shows that RDSCB 
is independent of the change of edge uniaxial anisotropy which is reflected in parallel colored 
stripes which is clear since the edge localized states change to center band states, which keep the 
RDSCB is nearly constant. The figure also shows that RDSCB increases with increasing the stripe 
width. This conclusion agrees with the result in Figure [10} 



3. The effects of edge exchange on zigzag stripe RDSFL and RDSCB 



Figure 12 shows the color contour plot for the effects of edge exchange and zigzag stripe width 
on its RDSFL. In these types of plots, the line with edge exchange equal to interior exchange, i.e. 
J e = 1, represents the calibration line of the figure. Here, this line is no more than the Figure [9] It 
is clear that RDSFL is nearly independent of the value of edge exchange. The main change on the 



RDSFL is when J e = 0. Figure 13 shows the effect of edge exchange at J e = and zigzag stripe 
width on its RDSFL, which by comparing it with Figure [9] shows the large increase in RDSFL 
values for both odd and even stripes width as the effect of edge exchange at J e = 0. The behavior 
of odd width stripes does not change with edge exchange at J e = 0. The behavior of even width 

17 




N number of rows 

FIG. 12. The effects of edge exchange and zigzag stripe width on its RDSFL 



Edge exchange J =0 




n 1 1 ' 1 1 1 ' r 

10 20 30 40 50 

N number of rows 



FIG. 13. The effect of edge exchange at J e = and zigzag stripe width on its RDSFL 



18 



stripes does change with edge exchange at J e = 0, first in the RDSFL become constant for small 
even width stripes at about N = 2S the RDSFL has fast increase to become just above the RDSFL 
odd width stripes RDSFL, which show that the probability of tunneling of edge localized states 
depend on both the even stripe width and the edge exchange. 




10 20 30 40 50 

N number of rows 



FIG. 14. The effects of edge exchange and zigzag stripe width on its RDSCB 



Figure 14 shows the color contour plot for the effects of edge exchange and zigzag stripe width 
on its RDSCB. It is clear from the figure that RDSCB decreases with an increase of the exchange. 
This decrease is particularly large in edge exchange range from 0.5 to 1.0 which shown as a 
curvature in the colored RDSCB stripes, while the RDSCB increases when increasing the stripe 
width. Out of edge exchange range 0.5 to 1.0 the RDSCB depends mainly on stripes width which 
is reflected in parallel colored stripes. 



4. The effects of impurities on zigzag stripe RDSFL and RDSCB 

The study of magnetic impurities effects on zigzag stripe is important for expected applica- 
tions. In this section the results are represented for the effects of introducing one and two lines of 
magnetic impurities on zigzag 20 and 21 width stripes on their RDSFL and RDSCB. 

There are two parameters for the impurities that engineering the energy band for magnetic 

19 



zigzag stripes: The first one is the strength of magnetic interaction represented by line of impurity 
exchange 7/ between the impurities line and the stripe materials [|22l . which here take the range 
of values from to 2 in the units of stripe materials magnetic exchange 7. The second parameter 
is the impurities line position, which can take the value from second to one line before the stripe 
end, the line position is alternative between even position number in sublattice B and odd position 
number in sublattice A (see Figure [T]). 

The first case to be shown here is the introduction of one line of magnetic impurities on zigzag 
20 width stripe. The results for the effects of the position of one impurities line with impurities 
exchange 7/ from to 2 for N = 20 zigzag stripe show that RDSFL is nearly independent of the 
position of impurities line and it is impurities exchange value except when 7/ = which is similar 
to the case of edge exchange effect. 



Figure 15 shows the effect of the impurities line position for impurities exchange at 7/ = on 
RDSFL of zigzag 20 width stripe. RDSFL is dependent on the parity of impurities line position. 
The RDSFL for even positions of impurities line begins high and decreases stepwise and beginning 
from position 12 become constant until position 18. The RDSFL for odd positions of impurities 
line have opposite behaviors; it begins small constant and beginning from position 11 begin to 
increases stepwise. 

Since edge localized states at Fermi level depend mainly on the edge geometry and the width 
of zigzag stripes, we expected that above behavior is related to the geometries and the widths of 
zigzag sub stripes and their interaction. The geometries of zigzag sub stripes as follows for even 
positions the impurities line is in the sublattice B and subdivide the stripe to one odd stripe type 
A (i.e. begin and end with sublattice A) and one even stripe begin with sublattice A and end 
with sublattice B. As the even position increases, the odd stripe A increases and the even stripe 
decreases. It is clear that odd stripe contribution is higher in edge localized. This is because of its 
edge has 1 coordination number, while the even stripe has both edges with 2 coordination number. 

For odd positions the impurities line is in the sublattice A and subdivide the stripe to one even 
stripe begin with sublattice A and end with sublattice B and odd stripe type B (i.e. begin and 
end with sublattice B). As the odd position increase the even stripe increases, and the odd stripe 
decreases. It is therefore clear that odd stripe contribution is higher in edge localized. This occurs 
when one of its edge has 1 coordination number, while the even stripe has both edges with 2 
coordination number. 



Figure 16 shows the effects of one line of impurities position and impurities exchange for 

20 



460- 



450- 



CO 
Q 
DC 

430-1 



420- 



410 



Line of impurity exchange J =0 



Odd Number N 




Zigzag N=20 



4 6 8 10 12 14 16 

The position of the impurity line 



- 1 — 1 — r~ 

18 20 



FIG. 15. The effect of one line of impurities position with impurities exchange Jj = for N = 20 zigzag 
stripe on its RDSFL 




6 8 10 12 14 

The position of the impurity line 



FIG. 16. The effects of one line of impurities position and impurities exchange for N = 20 zigzag stripe on 
its RDSCB 



21 



N = 20 zigzag stripe on its RDSCB, beginning from impurities exchange with value 1.0 which 
means no impurities. As the impurities exchange increases, the RDSCB decreases. As impurities 
exchange reach 1.4 the RDSCB decreases more at the line of impurities positions 9 and 12. As 
impurities exchange from 1.6 to 2, the RDSCB decreases more at the line of impurities positions 5, 
8, 9, 12, 13 and 16. As impurities exchange decreases than 1.0, the RDSCB increases as impurities 
exchange reach 0.4 the RDSCB increases more at the line of impurities positions 3, 6, 7, 10, 11, 
14, and 15. As impurities exchange from 0.4 to 0.0, the RDSCB has peak values for impurities 
line positions 2 and 19. 

The addition of second impurities line to the zigzag stripe increases the possibility to tune 
the magnetic properties of the stripes to suite more expected technological devices applications. 
Figure s [T7| and [T7| show the effects of second line of impurities position and its impurities exchange 
on RDSFL and RDSCB of N = 20 zigzag stripe. 

The results for the effects of second impurities line position with impurities exchange J It from 
to 2 for N = 20 zigzag stripe with first line of impurities at position 1 1 with impurities exchange 
7/ = show that RDSFL is nearly independent on the position of second impurities line and its 
impurities exchange value except when J n = which is similar to the cases of one impurities line 
and edge exchange effect. 



Figure 17 shows the effect of second impurities line position with impurities exchange Jn = 
for N = 20 zigzag stripe with one line of impurities at N = 11 with impurities exchange 7/ = 
on its RDSFL. The addition of second line increases the RDSFL more than one line of impurities 
and the behavior change since the new sub stirpes edges geometries and their interaction change. 
The RDSFL becomes nearly constant and independent on the second line position parity until the 
first impurities line position. The second line superimposed on the first line left only the effect of 
first line on the RDSFL. After this position the even position number switch to low nearly constant 
value while odd position number switch back to high nearly constant value. 



Figure 18 shows the effect of second impurities line position with impurities exchange Jn = 
for N = 20 zigzag stripe with one line of impurities at N = 11 with impurities exchange 7/ = 
on its RDSCB, beginning from impurities exchange with value 1.0 which means no impurities. 
As impurities exchange increases, the RDSCB decreases and as impurities exchange reach 1.4, 
the RDSCB decreases more at certain second impurities line positions. As impurities exchange 
from 1.8 to 2, the RDSCB decreases more at most second impurities line positions. As impurities 
exchange decreases than 1.0, the RDSCB increases gradually in most positions. As second impu- 

22 



900-, 



800- 



700- 



q 600- 



500- 



400- 



Line of impurity exchange J =0 




Odd Number N 



Zigzag 

N=20 

Nl=11 



Even Number N 



-i — | — i — | — i — | — i — | — i — | — i — | — i — | — i — | — i — | — i — | — i 



2 4 6 8 10 12 14 16 18 

The position of the second impurity line 



20 



FIG. 17. The effect of second line of impurities position with impurities exchange Jjj = for N = 20 zigzag 
stripe with one line of impurities @ N = 11 with impurities exchange Jj = on its RDSFL 



Zigzag 
N=20 
Nl=11 

RDSCB 




T 

6 8 10 12 14 16 1? 

The position of the second impurity line 



FIG. 18. The effect of second line of impurities position and impurities exchange for N = 20 zigzag stripe 
with one line of impurities @ TV = 1 1 with impurities exchange Jj = on its RDSCB 



23 



rities line positions 2 and 9 there are especial increase in RDSCB at impurities exchange from 0.4 
to 0.0. 

At first impurities line position, the second line superimposed on the first line left only the 
effect of first line on the RDSCB. The second impurities line position 14 have especial RDSCB 
behavior due to the existence of stripe with 2 lines between two zero exchange lines creating 
accumulation sites for magnons which create two localized states: one in conduction band and 
the other in valance band displayed as apeak RDSCB at second impurities exchange equal to 
as the impurities exchange increases the accumulation sites decreased very fast to remove the two 
localized states and drop RDSCB to the value of second impurities exchange equal to one. 

The results for the effects of one impurities line position with impurities exchange Jj from 
to 2 for N = 21 zigzag stripe show that RDSFL is nearly independent on the position of the 
impurities line and its impurities exchange value except when Jj = which is similar to the cases 
of one impurities line and edge exchange effect. 



Figure 19 shows the effect of the impurities line position for impurities exchange at Jj = 
on RDSFL of zigzag 21 width stripe. RDSFL is dependent on the parity of impurities line posi- 
tion. The RDSFL for even positions of impurities line is high and constant independent on even 
lines positions which is very similar to general zigzag odd stripes behavior. While the RDSFL 
for odd positions of impurities line have revers behavior it begins small constant and beginning 
from position 9 begin to increases slowly stepwise which is similar to general zigzag even stripes 
behavior. 

The above behavior is related to the geometries and the widths of zigzag sub stripes and their 
interaction in this case. The geometries of zigzag sub stripes as follow for even positions the 
impurities line is in the sublattice B and subdivide the stripe to two odd stripe type A (i.e. begin 
and end with sublattice A). As the even position increases the first odd stripe A increases and the 
other odd stripe A decreases. It is clear that both odd stripe contribution is equal in edge localized 
states since one of their edges has 1 coordination number, which result very similar behavior to 
general zigzag odd stripes. 

While for odd positions the impurities line is in the sublattice A and subdivide the stripe to one 
even stripe begin with sublattice A and end with sublattice B and even stripe too but begin with 
sublattice B and end with sublattice A. As the odd position increases first even stripe increases 
and the other even stripe decreases, the interaction between the two even stripes gives very similar 
behavior to general zigzag even stripes. 

24 



CO 

a 



600 4 



500- 



400 



300- 



Even Number N 



Line of impurity exchange J =0 



Odd Number N 



Zigzag N=21 



6 8 10 12 14 16 

The position of the impurity line 



18 



20 



FIG. 19. The effect of one line of impurities position with impurities exchange Jj = for N = 21 zigzag 
stripe on its RDSFL 



Zigzag 
N=21 
RDSCB 




6 8 10 12 14 16 1S 
The position of the impurity line 



FIG. 20. The effect of one line of impurities position and impurities exchange for N = 21 zigzag stripe on 
its RDSCB 



25 



Figure 20 shows the effects of one line of impurities position and impurities exchange for N = 
21 zigzag stripe on its RDSCB, beginning from impurities exchange with value 1 .0 which mean no 
impurities, as impurities exchange increases the RDSCB decreases as impurities exchange reach 
1.2 the RDSCB decreases more at all lines of impurities positions except at position 19. 

As impurities exchange decreases than 1.0, the RDSCB increases gradually in most position. 
At impurities line positions 2 and 17 there are especial increases in RDSCB at impurities exchange 
from 0.2 to 0.0. Impurities line position 19 have special RDSCB behavior due to the existence of 
stripe with 2 lines between one zero exchange line and the edge creating accumulation sites for 
magnons. This creates two localized states: one in conduction band and the other in valance band 
displayed as apeak RDSCB at impurities exchange equal to as the impurities exchange increases 
the accumulation sites decreased very fast to remove the two localized states and drop RDSCB to 
the value of second impurities exchange equal to one. 

The results for the effects of second impurities line position with impurities exchange /// from 
to 2 for N = 21 zigzag stripe with first line of impurities at position 1 1 with impurities exchange 
7/ = show that RDSFL is nearly independent on the position of second impurities line and its 
impurities exchange value except when J n = which is similar to many cases before. 

Figure [2T] shows the effect of second impurities line position with impurities exchange Jn = 
for N = 2\ zigzag stripe with one line of impurities at N = 11 with impurities exchange Jj = on 
its RDSFL. The addition of second line increase the RDSFL more than one line of impurities and 
the behavior change since the new sub stirpes edges geometries and their interaction change. The 
RDSFL become nearly constant and independent on the second line position parity until the first 
impurities line position, and the second line superimposed on the first line left only the effect of 
first line on the RDSFL. After this position the even position number keep nearly constant value 
while odd position number switch back to high nearly constant value. 



Figure 22 shows the effect of second impurities line position with impurities exchange J n for 
N = 21 zigzag stripe with one line of impurities at N = 11 with impurities exchange Jj = on 
its RDSCB, beginning from impurities exchange with value 1.0 which means no impurities. As 
impurities exchange increases the RDSCB decreases as impurities exchange reach 1 .6 the RDSCB 
decreases more at certain second impurities line positions. As impurities exchange decreases than 
1.0, the RDSCB increases gradually in most positions. At second impurities line positions 2, 9, 
16, and 17 there are special increases in RDSCB as impurities exchange from 0.4 to 0.0. 

At first impurities line position, the second line superimposed on the first line left only the effect 

26 



700 -, Line of impurity exchange J =0 



600- 



500- 



w 

Q 

^ 400-1 



300- 



200- 



Even Number N 




Zigzag 

N=21 

Nl=11 



Odd Number N 



n — 1 — r 



t — 1 — r 



-i 1 r 



i — 1 — r 



-i 1 r- 



2 4 6 8 10 12 14 16 18 20 

The position of the second impurity line 



FIG. 21. The effect of second line of impurities position with impurities exchange Jjj = for Af = 21 zigzag 
stripe with one line of impurities @ N = 11 with impurities exchange 7/ = on its RDSFL 




The position of the second impurity line 



FIG. 22. The effect of second line of impurities position and impurities exchange for N = 20 zigzag stripe 
with one line of impurities @ Af = 1 1 with impurities exchange Jj - on its RDSCB 



27 



of first line on the RDSCB. The second impurities line positions 14 and 19 have special RDSCB 
behavior due to the existence of stripe with 2 lines between two zero exchange lines creating 
accumulation sites for magnons. This creates two localized states: one in conduction band and 
the other in valance band displayed as a peak RDSCB at second impurities exchange equal to 
as the impurities exchange increases the accumulation sites decreased very fast to remove the two 
localized states and drop RDSCB to the value of second impurities exchange equal to one. 



B. Armchair stripes results 



Figures 23 show the dispersion relations for three armchair stripes with widths 20, 21, and 22. 
Where the nearest neighbor exchange 7, 7 has a constant value J through all the stripe including 
the stripe edges, the same for uniaxial anisotropy term D,, the obtained dispersion relation is very 
near to the obtained dispersion relation for armchair graphene ribbons with same size [|22l . only 
the dispersion curves here are shifted due to a effect. 

It is clear that the shape of the dispersion relations for armchair stipes depends on the stripe 
width. In general, the minimum of conduction band and the maximum of valence band are located 
at q x = for each stripe. In 20 lines stripe they touch each other at the Dirac point, while for 21 
and 22 lines stripes they have two different types band gaps. This behavior is a famous behavior 
seen in graphene armchair ribbons [|22l[34ll . This repeated pattern of the dispersion relations for 
armchair stripes can be described mathematically as periodic function in the number of lines as 3 i 
and 3/+ 1 for stripe with band gaps while 3i + 2 for gap less stripes where i = 1, 2, 3, • • • . The origin 
of this behavior is understood as consequence of topologically ladder system nature for armchair 
geometry, i.e. here cyclic chains with interchain hopping QTl [32l [341 - 1381 . 

In armchair geometry the sites from sublattice A are in the same line with sites from sublattice 
B, which is not the case in zigzag stripes. This removes the symmetry between adjusted lines and 
eliminates the degeneracy in armchair stripes without impurities [22]. There is no localized edge 
states show up in armchair without impurities as in case of graphene [|22l[3Tll32l . 



Figures 29 show the modified dispersion relations due to the effect of introducing substitutional 
a magnetic impurities line at row 1 1 of the armchair stripes with 20, 21, and 22 lines. Again as in 
the case of magnetic zigzag stripe, the new dispersions with impurities line for armchair magnetic 
stripes show exactly the same behavior seen in the same case for armchair graphene ribbons [|22l . 
but shifted in the case of magnetic stripes due the a effect. The introduction of the impurities line 

28 




29 



■ ' 1 1 1 

-0.5 0.0 0.5 

q a/it 




-2- 

' 1 ' 1 

-0.5 0.0 0.5 

q x a/jt 



FIG. 24. Spin waves dispersion for armchair 2D Heisenberg ferromagnetic honeycomb stripes with an 
impurity line at line number 11, where J - J e - \, Jj - 0.07 D - D e = Dj - 1.0 and a - 1.01 for 
N = 20,21 and 22 respectively. 



30 



have the effect as the case of zigzag stripes which is splitting the stripe to two interacted stripes 
with different sizes. In case of 20 line stripe the new stripes are 10 lines and 9 lines, in case of 21 
line stripe the new two stripes each 10 lines which lead to completely degenerate dispersion, and in 
case of 22 line stripe the new two stripes new stripes are 10 lines and 1 1 lines. The strength of the 
interaction between the two sub stripes depends on the value of the impurities exchange value 7/. 
The figures show case when 7/ = 0, in this case the expanded edge localized states in Fermi level 
are appear as the case of zigzag stripes. Those localized states are understood as accumulation 
sites for magnons in the interface created by the tunneling between the two substripes through 
the impurities line, as the distance between the two substripes increase the localized states density 
decreases until the total dispersion for the system show non-interacted individual dispersions for 
the two substripes without any localized states, while the situation is different in the case of zigzag 
stripes where there are intrinsic edge localized states beside the one due to the impurities line 
tunneling interface. 



Figures 25 show the modified dispersion relations due to the effect of introducing substitu- 
tional a magnetic impurities lines at rows 11 and 14 of the armchair stripes with 20, 21 and 22 
lines. Again, the introducing of the impurities lines have the effect of splitting the stripe to three 
interacted stripes with different sizes, in case of 20 line stripe the new stripes are 10 lines, 2 lines 
and 6 lines, in case of 21 line stripe the new two stripes 10 lines, 2 lines and 7 lines, and in case 
of 22 line stripe the new two stripes 10 lines, 2 lines and 8 lines. The existence of stripe with 
2 line which is stripe of type 3i+2 where i=0, force the band gapless feature in the three stripe. 
The absence of spacial phenomenon like intrinsic localized edge states and accumulation two line 
stripe in armchair stripes reflect the importance of stripe topology as armchair or zigzag in the 
follow of the nearest neighbors exchange inside the stripe which will be studied later in this work. 



As we see from the results above, that armchair type stripes have band gaps while there is no 
intrinsic localized edge state, and for the applications of similar armchair graphene nanoribbons 
the energy band gap is very important [|2Tll22l[33l . Therefore RDSCB and the change in the band 
gap are the good parameters for armchair stripes to study the effects of edges and impurities on 
their dispersions relations. 

31 



-0.5 0.0 0.5 

q ain 




-2- 

1 1 1 1 

-0.5 0.0 0.5 

q a/n 



FIG. 25. Spin waves dispersion for armchair 2D Heisenberg ferromagnetic honeycomb stripes with an 
impurity lines at line number 11 and line number 14, where J = J e - \, Jj = Jjj = 0.0/ D - D e - Dj - 1.0 
and a = 1.01 for N = 20, 21 and 22 respectively. 



32 



1. The effect of armchair stripe width on its RDSCB 



Figure 26 shows the effect of armchair stripe width on its RDSCB. The RDSCB is highly 
dependent on the armchair stripe width type as 3i, 3i+l, and 3i+2. There is a repeated pattern 
between the stripe width types and their RDSCB as follow the sequence of first step is 3i+l, 3i, 
3i+2, 3i+l, and 3i+2 which then repeated as the width increases. Overall, the average value of the 
RDSCB linearly increases with increasing stripe width. 



5000- 



4000- 



g 3000- 

w 

Q 

DC 

2000-1 



1000- 



V 4 



10 20 30 

N number of rows 



■ 


3i 


• 


3i+1 


▲ 


3i+2 



—\— 

40 



I 

50 



FIG. 26. The effect of armchair stripe width for the three armchair types on the relative density of states of 
center band (RDSCB) 



2. The effects of edge uniaxial anisotropy on armchair stripe RDSCB 

Figure [27] shows the color contour plot for the effects of edge uniaxial anisotropy and armchair 
stripe width on its RDSCB. The Figure shows that RDSCB is nearly independent on the change of 
edge uniaxial anisotropy which as the case of zigzag stripe, is reflected in parallel colored stripes, 
except at D e = D where there are certain stripe widths with higher value RDCSB than surrounding 
numbers, which is first time to be seen in armchair stripes. Those numbers are 9, 15, 21, 27, 33, 
39, and 45 which of type 3z for which i odd primary number 3, 5, 7, 9, 11, 13 and 15. They are 

33 



also clear in Figure [26] as the most outer 3i stripes widths (black squares) which show that at those 
lengths the system wave function has high density of states at the center band [34]. In general, 
changing the edges insite energy breaks the symmetry of the dispersion relation and move slightly 
the Fermi Level which also seen in graphene nanoribbon [35], but nearly has no effect on the 
RDSCB value. It is clear that RDSCB is increasing with increasing the stripe width which agree 



with result in figure 26 




N number of rows 



FIG. 27. The effects of edge uniaxial anisotropy and armchair stripe width on its RDSCB 



3. The effects of edge exchange on armchair stripe RDSCB 



Figure 28 shows the color contour plot for the effects of edge exchange and zigzag stripe width 
on its RDSCB. It is clear from the figure that RDSCB is decreasing with the increasing of the edge 
exchange which is shown as a curvature in the colored RDSCB stripes where the zigzags lines at 
the boundary of the colored stripes come from armchair types depends. The RDSCB is increasing 
with increasing the stripes width which agree with with result in figure [26} 

34 




RDSCB 



N number of rows 



FIG. 28. The effects of edge exchange and zigzag stripe width on its RDSCB 



4. The effects of impurities on armchair stripe RDSCB 



The study of magnetic impurities effects on armchair stripe is important for expected appli- 
cations as in case of zigzag stripes. In this section the results are represented for the effects of 
introducing one and two lines of magnetic impurities on armchair 20, 21 and 22 width stripes on 
their RDSCB. 

As in case of the zigzag case, there are two parameters for the impurities that engineering the 
energy band for magnetic armchair stripes: The first one is the strength of magnetic interaction 
represented by line of impurity exchange 7/ between the impurities line and the stripe materials 
j|22|. which here take the range of values from to 2 in the units of stripe materials magnetic 
exchange J. The second parameter is the impurities line position, which can take the value from 
second to one line before the stripe end, the line position is alternative between even position 
number. 



Figure 29 shows the effects of one line of impurities position and impurities exchange for 
N = 20 and N = 22 armchair stripes on their RDSCB. The figure shows that the behavior of 
the RDSCB is nearly the same for both even width stripes, beginning from impurities exchange 
with value 1 .0 which means no impurities, as impurities exchange increases the RDSCB decreases 



35 



and it is independent on the impurity line position. As impurities exchange decreases than 1.0, 
the RDSCB increases as impurities exchange reach 0.0 the RDSCB has maximum value and it is 
independent on the impurity line position. This is due to the creation of expanded edge localized 
states in Fermi level, the existence of one line impurity divide the armchair 20 and 22 width stripes 
to two interacting substripes one odd and one even with no probability to have symmetry between 
them, which make this position independent. 



Figure 29 shows the effects of one line of impurities position and impurities exchange for 
N = 21 armchair stripe on its RDSCB, beginning from impurities exchange with value 1.0 which 
means no impurities, as impurities exchange increases the RDSCB decreases periodically with 
maximum decreasing in odd impurity line positions. As impurities exchange decreases than 1.0, 
the RDSCB increases as impurities exchange reach 0.2 the RDSCB has maximum value in even 
impurity line positions. The behavior can be explained by the fact that the existence of one line 
impurity divide the armchair 21 width stripe to two interacting substripes with their width parities 
depends on the impurity line position, in case of even impurity line position the two substripes 
are one odd and one even which is same of armchair stripe with 20 width, while in case of odd 
impurity line position the two substripes are even, which increase the symmetry and degeneracy 
between the two substripes and reduce the value of RDSCB. 



Figure 30 shows the effect of second impurities line position with impurities exchange J n for 
N = 20 and N = 22 armchair stripes with one line of impurities at N = 11 with impurities 
exchange 7/ = on its RDSCB. The figures show that the behavior of the RDSCB is nearly the 
same for both even width stripes as the RDSCB is dependent on the position of second impurities 
line with respect to the position of first impurities line, which is explained by the fact that the first 
impurities line divide the armchair 20 and 22 stripes to two substripes, the first one is even stripe 
with 10 width for both 20 and 22 stripes and the second one is odd stripe with 9 and 11 widths 
for stripes. When the second line change its position in the even substripes, the RDSCB shows 
the same behavior of one line impurities in even armchair stripes. At first impurities line position, 
the second line superimposed on the first line left only the effect of first line on the RDSCB. 
As the second line change its position in the odd substripes, the RDSCB shows similar behavior 
of one line impurities in odd armchair stripes with shift for the RDSCB peaks form even second 
impurities line positions to odd second impurities line positions at second impurities line exchange 
equal to zero due to the interaction with first even substripe with 10 width. 

Figure [30] shows the effect of second impurities line position with impurities exchange J u for 

36 




The position of the impurities line 




2 4 6 8 10 12 14 16 18 20 

The position of the impurities line 




The position of the impurities line 

FIG. 29. The effect of one line of impurities position and impurities exchange on armchair stripe RDSCB 
for stripes width N = 20, 21 and 22 respectively. 



37 



N = 21 armchair stripes with one line of impurities at N = 11 with impurities exchange Jj = 
on its RDSCB, beginning from impurities exchange with value 1.0 which means no second impu- 
rities, as impurities exchange increases, the RDSCB decreases independently on second impurity 
line positions. As impurities exchange decreases than 1.0, the RDSCB increases as impurities 
exchange reach 0.0 the RDSCB has maximum value in even impurity line positions. 



5. The effect of armchair stripe width on its energy band gap 

The direct energy band gab seen at q x = above in armchair stripes is similar to graphene 
armchair ones and it is very important from applications point of view ll2Tll22l[33l . The relation 
between the graphene armchair nanoribbons types, its width and its band gab has been studied 
extensively both experimentally |39l and theoretical OH |32l [35J |40l |4TJ. It has been found that 
the main factor in the armchair nanoribons energy band gab of types 3i and 3i + 1 behavior is 
the quantum confinement which reflected in the inverse dependance of energy band gab E g on the 
stripe width W such that E g ~ W' 1 . 

We used our model to study the variation of bandgaps of the three types magnetic armchair 



stripes as a function of width (number of rows) the results are shown in Figure 3 1 The Figure 
shows that the energy band gap E g armchair types 3i and 3?'+ 1 have the same width W dependance 
seen in graphene armchair nanoribons, i.e. E g ~ W~ l , which show the close similarity between 
graphene nanoribons and magnetic stripes. 



6. The effects of edge exchange on armchair stripe energy band gap 

In our study of width effect on the energy band gab of magnetic armchair stripe, we assumed 
that the edges sites have the same exchange as interior sites, i.e. J e = J, but this is not the case, 
as edges sites have different coordination number consequently their exchange is different from 
interior, also there is a technical possibility to engineering their magnetic properties and as it is 
found in armchair graphene nanoribbons case that the edges play important rule in their energy 
band gap [|40l . the same is expected for magnetic stripes. Figure [32] shows the behavior of spin 
wave energy modes at direct band gap point, i.e. q x = with the variation of the edges exchange 
strength J e . Figure [32] (a) shows the behavior of all modes for a 20-line armchair stripe with the 
variation of the edges exchange strength, as edges exchange increases some modes begin to bend 

38 




2 4 6 8 10 12 14 16 18 20 

The position of the second impurities line 




The position of the second impurities line 

FIG. 30. The effect of second line of impurities position and impurities exchange on armchair stripe RDSCB 
with one line of impurities @ N = 1 1 with impurities exchange Jj = for stripes width N = 20, 21 and 22 
respectively. 



39 



0.9 i 
0.8- 
0.7- 
0.6- 



-3 0.5- 
C/) 



LU 



0.4- 
0.3- 
0.2 
0.1 - 
0.0- 







• 


3i 


■ 


3i+1 


▲ 


3i+2 



10 



20 30 40 50 

Armchair stripe width 



60 



I 

70 



FIG. 31. The variation of bandgaps of the three types armchair stripes as a function of width (number of 
rows) W 

and crossing other modes some of them leave all the stripe energy band, the over all behavior is 
the same for 21 and 22 stripes. 

Figure [32] (b), (c), and (d) show the amplified region around (co/S J) = a for 20, 21 and 22 stripe 
widths respectively. They show similar behavior beginning at zero edge exchange the minimum 
conduction band mode and the maximum valance band meet together and consequently there is 
no band gap which shown as a localized states at Fermi level (co/S J) = a, as the edge exchange 
increases the two mode splitting producing a band gap for the stripes widths which then crossing 
each other at certain value of edge exchange depending on the stripe type. In general as the edge 
exchange approach the value of interior sites exchange, the modes rearranged to show the behavior 
seen before in Figure ??. 

Figure [33] shows the variation of the energy gap, which is the difference between the minimum 
conduction band mode and the maximum valance band mode in the stripe, against the strength of 
the edge exchange for 20, 21 and 22 armchair stripes. The three energy band gap starts from zero 
and increases to reach maximum and then decreasing to reach zero again at edge exchange value 
depending on the stripe width, which are 0.5 for 21 stripe, 1.0 for 20 stripe and 2.0 for 22 stripe. 
After this minimum each energy band gap starts to increases again, the 20 stripe energy band gap 

40 



1 2 

J /J J /J 

e e 




J /J J /J 

e e 



FIG. 32. g r = dispersion of the modes (a) 20-line armchair stripe as a function of the strength of the edge 
exchange J e which amplified in the region around (a>/S J) = a in (b). For (c) and (d) the amplified lower 
energy region for N=21 and 22 respectively 

increases to seam constant value, while the 21 stripe energy band gap increase to new maximum 
value and the decreases, and the 22 stripe energy band gap decreases slowly. 



7. The effects of impurities on armchair stripe energy band gap 

Introducing the impurities in armchair stripes modify their energy bands as seen in Figures 



25 and 30 The value of impurities exchange there chosen to be equal to zero, but as impurities 



exchange increases from zero the band gap behavior should be similar to graphene case Q22 



Figure 34 shows the behavior of spin wave energy modes at direct band gap point, i.e. q x = with 
the variation of the impurities exchange strength 7/, where the impurities line in the 1 1th line. The 
Figure|34](a) shows the behavior of all modes for a 20-line armchair stripe with the variation of the 

41 



0.35 n 



0.30- 



w 

c 

LU 



0.20- 




0.15- 



J /J 



FIG. 33. Variation of the energy gap against the strength of the edge exchange for an armchair stripes. Solid 
line for a 20- line ribbon, dashed line for 21 lines, and dotted line for a stripe with 22 lines. 

impurities exchange strength. As impurities exchange increases, some modes begin to bend and 
crossing other modes some of them leave all the stripe energy band, the over all behavior is the 
same for 21 and 22 stripes, which is the same behavior seen before in edge exchange and graphene 
case [|22l. 



Figure 34 (b), (c), and (d) show the amplified region around (to/S J) = a for 20, 21 and 22 
stripe widths respectively. They show similar behavior beginning at zero impurities exchange the 
minimum conduction band mode and the maximum valance band meet together. Consequently, 
there is no band gap which shown as a localized states at Fermi level (co/S J) = a, as the impurities 
exchange increases the two mode splitting producing a band gap for the stripes widths which 
then crossing each other at certain value of impurities exchange depending on the stripe type. 
In general, as the impurities exchange approach the value of interior sites exchange, the modes 
rearranged to show the behavior seen before in edge exchange case. 



Figure 35 shows, as in the case of edge exchange, the variation of the energy gap against the 
strength of the impurities exchange for 20, 21 and 22 armchair stripes. The three energy band 
gap starts from zero and increases to reach maximum, while 22 stripe begin to decreases slowly to 
nearly constant value, and the 20 and 21 stripe begin to decreasing to reach zero again at impurities 



42 




J/J J/J 



FIG. 34. q x = dispersion of the modes (a) 20-line armchair stripe with an impurities line in the 11th line 
as a function of the strength of the impurity exchange Jj, which amplified in the region around (a>/S J) - a 
in (b). For (c) and (d) the amplified lower energy region for N=21 and 22, respectively 

exchange value depending on the stripe width, which are 0.5 for 21 stripe, 1.0 for 20 stripe. After 
this minimum for the two stripes each energy band gap starts to increases again, while the 20 stripe 
energy band gap increase slowly to seam constant value, the 21 stripe energy band gap increase 
faster to nearly constant. 

Figures [36] show color contour plot of the behavior of energy band gap for the three stripes as 
a function in both impurities lines positions and their impurity exchange strength. The right part 
shows a periodic variation in band gap of the three stripes with changing both the position and 
the exchange strength of one impurity line, also the variation is symmetric about the stripe center 
which is the same behavior found in graphene [|22]|. The left part show the removing of the periodic 
variation in band gap for the three stripes while its symmetric about center is removed for the 20 
and 22 two stripe which is the same behavior for their RDSCB. The left part is due to adding a 

43 



0.35-1 
0.30- 




FIG. 35. Variation of the energy gap against the strength of the impurity exchange with impurities line in 
the 1 1th line of armchair stripes. Solid line for a 20-line stripe, dashed line for 21 lines, and dotted line for 
a stripe with 22 lines. The impurities are always in line number 1 1 . 

second impurity line to the three stripes while fixing the first impurity line in row number 1 1 with 
zero impurity exchange, which show the great ability to tune the energy band gap of armchair 
stripes with adding one or more magnetic impurity line. 

V. DISCUSSION AND CONCLUSIONS 

In this work, the second quantization form of Heisenberg Hamiltonian for ferromagnetic short 
range (a geometrical quantity) interaction between nanodots represented by nearest neighbor (NN) 
exchange is used to study the allowed spin waves modes, i.e. dispersion relations (energy 
band), for 2D Honeycomb Lattice (a geometrical quantity). 

The results of this study for ferromagnetic dots 2D honeycomb lattice stripes show almost 
coincidence with the results of graphene nanoribbons described by tight binding Hamiltonian for 
electronic short range (a geometrical quantity) interaction between carbons atoms represented by 
nearest neighbor (NN) hopping for 2D Honeycomb Lattice (a geometrical quantity). 

44 




2 4 6 8 10 12 14 16 18 2 4 6 8 10 12 14 16 18 

The position of the impurities line The position of the second impurities line 




2 4 6 8 10 12 14 16 18 20 2 4 6 8 10 12 14 16 18 20 

The position of the impurities line The position of the second impurities line 



FIG. 36. Color contour plot of the energy gap for the armchair stripes width 20, 21, and 22 left side showing 
the energy gap with the position of the first impurity line varying from line 2 to N-l. Right side showing 
the energy gap with the position of the second impurity line varying from line 2 to N-l, with first impurities 
line at row number 1 1 with Jj = 0. 



45 



From technological point of view, those results are very encouraging to fabricate a magnetic 
counterpart to graphene, which will lead to a new technology especially in the field of spintronic 
devices and magnetic applications. Also, the results show that many obtained researches results 
of graphene can be easily applicable to the magnetic case. 



ACKNOWLEDGMENTS 

This research has been supported by the Egyptian Ministry of Higher Education and Scientific 
Research (MZA). 



[1] M. Z. Ahmed, Study of electronic and magnetic excitations in the 2D materials represented by 
graphene and magnetic nano-ribbons, Ph.D. thesis, The University of Western Ontario (2011), |down-| 
load pdf version and a Viedo on Sciencestage.com. 



[2] O. V. Yazyev, Reports on Progress in Physics 73, 056501 (2010) 



[3] M. Mecklenburg and B. C. Regan, Phys. Rev. Lett. 106, 1 16803 (201 1) 



[4] R. S. E.O. Kamenetskii and M. Sigalov, (Mar 2003), arXiv:cond-mat/0303196vl [cond-mat.mes- 



hall] 



[5] M. Rontani, Nat Mater 10, 173 (2011) 



[6] O. Astafiev, K. Inomata, A. O. Niskanen, T. Yamamoto, Y. A. Pashkin, Y. Nakamura, and J. S. Tsai, 



Nature 449, 588 (2007) 



[7] R. Bratschitsch and A. Leitenstorfer, Nat Mater 5 , 855 (2 006). 
[8] M. A. Kastner, Physics Today 46, 24 (1993). 



[9] L. Malkinski, R. E. Camley, Z. Celinski, T. A. Winningham, S. G. Whipple, and K. Douglas, Journal 
|ofApplied P hysics 93, 7325 (2003)| 
[10] M. Tanaka, E. Saitoh, H. Miyajima, T. Yamaoka, and Y. Iye, Phys. Rev. B 73, 052411 (2006) 



[11] M. Stepanova and S. Dew, Nanofabrication: Techniques and Principles (Springer, 2011). 



[12] J. Richter, J. Schulenburg, and A. Honecker, in Quantum Magnetism, Lecture Notes in Physics, Vol. 
645, edited by U. Schollwck, J. Richter, D. Farnell, and R. Bishop (Springer Berlin / Heidelberg, 



46 



2004) pp. 85-153, 10. 1007/BFbOl 19592. 
[13] H. S. Darendelioglu, |Scripta Me tallurgica et Material ia~33, 1825 (1995)] 



[14] J. Oitmaa, C. J. Hamer, and Z. Weihong, Phys. Rev. B 45, 9834 (1992) 



[15] J. D. Reger, J. A. Riera, and A. P. Young, Journal of Physics: Condensed Matter 1, 1855 (1989) 



[16] A. A. Tsirlin, O. Janson, and H. Rosner, |Phys. Rev. B 82, 144416 (2010)| 



[17] E. Rastelli and A. Tassi, Phys. Rev. B 49, 9679 (1994) 



[18] M. T. Nguyen, Spin-wave excitations in ferromagnetic nanostructures, Ph.D. thesis, University of 
Western Ontario (2007). 

[19] H. T. Nguyen, Dipole-Exchange Spin Waves in Magnetic Nanomaterials , Phd, The University of West- 
ern Ontario (2009). 
[20] M.Ahmed, (2011), arXiv:1110.4369vl [cond-mat.mes-hall] 



[21] A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov, and A. K. Geim, Rev. Mod. Phys. 81, 
1 109 (2009)1 

[22] R. N. Costa Filho, G. A. Farias, and F. M. Peeters, |Phys. Rev. B 76, 193409 (2007)] 
[23] D. R. Bs, Quantum mechanics: a modern and concise introductory course (Springer, 2007). 
[24] R. L. Liboff, Introductory Quantum Mechanics, 1st ed. (Longman Higher Education, 1987). 
[25] L. Kantorovich, Quantum Theory of the Solid State: An Introduction (Springer, 2004). 
[26] U. Rssler, Solid State Theory: An Introduction (Springer, 2009). 

[27] K. F. Henrik Bruus, Many-body quantum theory in condensed matter physics: an introduction (Oxford 
University Press, 2004). 



[28] J. R. M. Karim M. Abadir, Matrix algebra, Vol. 1 (Cambridge University Press, 2005). 

[29] W. H. Press, Numerical recipes in FORTRAN :the art of scientific computing, Vol. 1 (Cambridge 

University Press, Cambridge England ; New York, 1992) p. 963. 
[30] C. Tao, L. Jiao, O. V. Yazyev, Y.-C. Chen, J. Feng, X. Zhang, R. B. Capaz, J. M. Tour, A. Zettl, S. G. 

Louie, H. Dai, and M. F. Crommie, Nat Phys advance online publication, (2011) 



[31] M. Fujita, K. Wakabayashi, K. Nakada, and K. Kusakabe, Journal of the Physical Society of Japan 



65, 1920 (1996) 



[32] K. Nakada, M. Fujita, G. Dresselhaus, and M. S. Dresselhaus, Phys. Rev. B 54, 17954 (1996) 



47 



[33] Z.-y. L. Bing Huang, Qi-min Yan and W. hui Duan, Frontiers of Physics in China 4, 269 (2009) 



[34] H. Zheng, Z. F. Wang, T. Luo, Q. W. Shi, and J. Chen, Phys. Rev. B 75, 165414 (2007) 



[35] M. Ezawa, Phys. Rev. B 73, 045432 (2006) 



[36] S. Gopalan, T. M. Rice, and M. Sigrist, |Phys. Rev. B 49, 8901 (1994) 



[37] M. Fabrizio, A. Parola, and E. Tosatti, |Phys. Rev. B 46, 3159 (19 92) 

[38] K. C.-M. O. H. Hosoya, H. Kumazaki and Y.-D. Gao, Pure Appl. Chem. 62, 445 (1990). 



[39] M. Y. Han, B. Ozyilmaz, Y. Zhang, and P. Kim, Phys. Rev. Lett. 98, 206805 (2007) 
[40] Y.-W. Son, M. L. Cohen, and S. G. Louie, |PhysTRev. Lett. 97, 216803 (2006)] 



[41] K. Wakabayashi, M. Fujita, H. Ajiki, and M. Sigrist, Phys. Rev. B 59, 8271 (1999) 



48